1.
ISWC 2010, Shanghai, 8th November, 2010
Ivan Herman (郝易文), W3C

2.
(2)
 For RDF people, it sounds very simple:
◦ RDFa is a serialization of RDF embedded in XHTML,
HTML, or XML in general

3.
(3)

4.
(4)

5.
(5)
 Apart from relational databases, most of the
data on the Web are in… (X)HTML content
 New content is generated every day
 How would one get structured data from that
information?

6.
(6)
 Do not generate RDF/XML files separately
◦ RDF/XML is complex
◦ it requires a separate storage, generation, etc
mechanism
 that is also valid for, e.g., Turtle
 even when authoring with, say, Emacs, creating an extra
file is a load

7.
(7)
 Add extra structured content to the (X)HTML
pages
 Let processors extract those and turn into RDF

8.
(8)
 Microformats
◦ reuses HTML attributes like @class, @title
◦ separate vocabularies (address, CV, …)
◦ difficult to mix microformats (no concept of
namespaces)
◦ possible to transform via, e.g., XSLT + GRDDL, but all
transformations are vocabulary dependent

9.
(9)
 Microdata
◦ adds new attributes to HTML5 to express metadata
◦ can use URI-s, it also fixes some vocabulary mappings
(e.g., to Dublin Core elements)
◦ works for simpler usages, not well suited for complex
vocabularies
◦ has no notion of datatypes, namespaces
◦ but generic processing becomes possible to generate
RDF

10.
(10)
 RDFa
◦ adds new (X)HTML/XML attributes
◦ has namespaces and URIs at its core; i.e., mixing
vocabulary is just as easy as in RDF
◦ complete flexibility for using Literals or URI Resources
◦ is a complete serialization of RDF
◦ generic processing becomes possible to generate RDF

11.
(11)

12.
(12)
 It is very important for RDF experts to
◦ know RDFa
◦ parse it alongside Turtle, RDF/XML or other
◦ when appropriate, generate RDFa pages

13.
(13)

14.
(14)
 RDFa means “RDF in attributes”. Ie:
◦ all RDF contents are defined through XML attributes
(no elements)
◦ the XML/HTML tree structure is used
◦ many of the attributes are defined by RDFa
 some attributes (@href, @rel) are also reused
◦ if possible, the text content is also reused (for literals)
as well as @href values

15.
(15)
 The same (X)HTML file:
◦ is used, unchanged, by browsers
 they ignore attributes they do not know
◦ can be used by specialized processors (or APIs) to
extract RDF triples

16.
(16)
 The current Recommendation is RDFa 1.0
 There is an RDFa1.1 in the making, almost
ready
 I will talk about RDFa1.1 and warn when the
feature is not available in RDFa1.0

17.
(17)
 Formally:
◦ RDFa WG defines Core and XHTML
◦ HTML WG defines HTML5
 this tutorial uses XHTML examples
XHTML
+RDFa
HTML5
+RDFa
SVG
1.2
…ODF
RDFa Core 1.1
(valid for any XML)

18.
(18)
 A browser usually asks for an HTML content:

19.
(19)
 Via content negotiations this goes to:

20.
(20)
 Via content negotiations this goes to:

21.
(21)
 But a client could ask for, say, Turtle:

22.
(22)
 The triples are embedded in the HTML file
◦ a client may know how to extract RDF triples directly
from that file; or
◦ an online “distiller” service is used; or
◦ the server is set up to generate the Turtle file
automatically

23.
(23)

24.
(24)
RewriteEngine On
RewriteBase /ns/entailment/data/
RewriteRule RDFS.ttl
/2007/08/pyRdfa/extract?format=turtle&
uri=http://www.w3.org/ns/entailment/data/RDFS.html [L]

25.
(25)

26.
(26)

27.
(27)

28.
(28)
<p about="http://www.w3.org/ns/entailment/RDFS"
property="http://purl.org/dc/terms/description">
Unique identifier for <em>RDFS Entailment</em>.</p>

29.
(29)
<p about="http://www.w3.org/ns/entailment/RDFS"
property="http://purl.org/dc/terms/description">
Unique identifier for <em>RDFS Entailment</em>.</p>
<http://www.w3.org/ns/entailment/RDFS>
… .

30.
(30)
<p about="http://www.w3.org/ns/entailment/RDFS"
property="http://purl.org/dc/terms/description">
Unique identifier for <em>RDFS Entailment</em>.</p>
<http://www.w3.org/ns/entailment/RDFS>
<http://purl.org/dc/terms/description>
… .

31.
(31)
<p about="http://www.w3.org/ns/entailment/RDFS"
property="http://purl.org/dc/terms/description">
Unique identifier for <em>RDFS Entailment</em>.</p>
<http://www.w3.org/ns/entailment/RDFS>
<http://purl.org/dc/terms/description>
"Unique identifier for RDFS Entailment." .

32.
(32)

33.
(33)
<a about="http://www.w3.org/ns/entailment/RDFS"
rel="http://www.w3.org/2000/01/rdf-schema#seeAlso"
href="http://www.w3.org/TR/2004/REC-rdf-mt-20040210/">
RDF Semantics.
</a>

34.
(34)
<a about="http://www.w3.org/ns/entailment/RDFS"
rel="http://www.w3.org/2000/01/rdf-schema#seeAlso"
href="http://www.w3.org/TR/2004/REC-rdf-mt-20040210/">
RDF Semantics.
</a>
<http://www.w3.org/ns/entailment/RDFS>
….

35.
(35)
<a about="http://www.w3.org/ns/entailment/RDFS"
rel="http://www.w3.org/2000/01/rdf-schema#seeAlso"
href="http://www.w3.org/TR/2004/REC-rdf-mt-20040210/">
RDF Semantics.
</a>
<http://www.w3.org/ns/entailment/RDFS>
<http://www.w3.org/2000/01/rdf-schema#seeAlso>
… .

36.
(36)
<a about="http://www.w3.org/ns/entailment/RDFS"
rel="http://www.w3.org/2000/01/rdf-schema#seeAlso"
href="http://www.w3.org/TR/2004/REC-rdf-mt-20040210/">
RDF Semantics.
</a>
<http://www.w3.org/ns/entailment/RDFS>
<http://www.w3.org/2000/01/rdf-schema#seeAlso>
<http://www.w3.org/TR/2004/REC-rdf-mt-20040210/> .

37.
(37)

38.
(38)
 The combination of @about with
@rel/@property and possibly @href covers most
of we need…
 But this is too complex for authors

39.
(39)
<http://www.w3.org/ns/entailment/RDFS>
<http://purl.org/dc/terms/description>
"Unique identifier for RDFS Entailment." .
<http://www.w3.org/ns/entailment/RDFS>
<http://www.w3.org/2000/01/rdf-schema#seeAlso>
<http://www.w3.org/TR/2004/REC-rdf-mt-20040210/> .
 with
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix dc: <http://purl.org/dc/terms/> .
<http://www.w3.org/ns/entailment/RDFS>
rdfs:seeAlso
<http://www.w3.org/TR/2004/REC-rdf-mt-20040210/> ;
dc:description "Unique identifier for RDFS Entailment." .

40.
(40)
 Use compact URI-s when possible
 Make use of XML structure for
◦ shared subjects
◦ shared predicates
◦ create blank nodes
◦ …

41.
(41)
 Just like in Turtle:
◦ define a prefix via @prefix
◦ use prefix:reference to abbreviate a URI

42.
(42)
<html>
…
<p about="http://www.w3.org/ns/entailment/RDFS"
property="http://purl.org/dc/terms/description">
Unique identifier for <em>RDFS Entailment</em>.</p>
…
</html>
 can be replaced by:
<html prefix="dc: http://purl.org/dc/terms/">
…
<p about="http://www.w3.org/ns/entailment/RDFS"
property="dc:description">
Unique identifier for <em>RDFS Entailment</em>.</p>
…
</html>

43.
(43)
 Can be anywhere in the XML tree and is valid for
the whole sub-tree
◦ i.e., the html element is not the only place to have it
 The same @prefix attribute can hold several
definitions:
◦ prefix="dc: http://purl.org… rdfs: http://…"

44.
(44)
 An alternative (deprecated) syntax is
◦ xmlns:dc="http://purl.org/dc/terms/"
 CURIEs and “real” URIs can be mixed
◦ if an attribute value can be interpreted as a CURIE, fine
◦ alternatively, it is considered as a URI
 CURIEs can be used on RDFa attributes only!
◦ e.g., not for @href

45.
(45)
 In RDFa 1.0
◦ only the xslt:XXX syntax is usable
◦ CURIEs on @about can only be used with the syntax:
about="[pref:ref]"
◦ Only CURIEs can be used on, e.g., @property or @rel
(no fallback on URIs)

46.
(46)
 The basic principle: @about is inherited by
children nodes
◦ i.e., no reason to repeat it

47.
(47)
<html prefix="dc: http://purl.org/dc/terms/
rdfs: http://www.w3.org/2000/01/rdf-schema#">
…
<body about="http://www.w3.org/ns/entailment/RDFS">
…
<p property="dc:description">
Unique identifier for <em>RDFS Entailment</em>.</p>
<p>…<a rel="rdfs:seeAlso"
href="http://www.w3.org/TR/2004/REC-rdf-mt-20040210">
RDFS Semantics</a>…</p>

48.
(48)
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix dc: <http://purl.org/dc/terms/> .
<http://www.w3.org/ns/entailment/RDFS>
rdfs:seeAlso
<http://www.w3.org/TR/2004/REC-rdf-mt-20040210/> ;
dc:description "Unique identifier for RDFS Entailment." .

49.
(49)

50.
(50)
<body about=".." prefix="dc: http://…"
<address>
<p property="dc:date">2010-07-05</p>
</address>
</body>
 This leads to:
@prefix dc: <http://…>
<..> dc:date "2010-07-05" .

51.
(51)
<body about=".." prefix="dc: http://… xsd: http://…"
<address>
<p property="dc:date" datatype="xsd:date">2010-07-05</p>
</address>
</body>
 This leads to:
@prefix dc: <http://…>
@prefix xsd: <http://…>
<..> dc:date "2010-07-05"^^xsd:date .

52.
(52)
 The basic rule says: the (RDF) Literal is the
enclosed text from the HTML content
 This is fine in 80% of the cases, but…
 It may not be natural in all cases! E.g.,
◦ 2010-07-05 is the “official” ISO format (for xsd:date)
◦ but “July 5, 2010” looks much more natural for a
human…

53.
(53)
<body about=".." prefix="dc: http://… xsd: http://…"
<address>
<p property="dc:date" datatype="xsd:date"
content="2010-07-05">July 5, 2010</p>
</address>
</body>
 Also leads to:
@prefix dc: <http://…>
@prefix xsd: <http://…>
<..> dc:date "2010-07-05"^^xsd:date .

54.
(54)

55.
(55)
 What we said is:
◦ @about sets the subject
◦ @href sets the object
 But that is not always good enough
◦ we do not always want active links (i.e., the "a"
element)
◦ what about other links in HTML?

56.
(56)
<body about="…">
…<img rel="foaf:depiction"
src="http://www.ex.org/img.png"/>…
</body>
 yields:
<…> foaf:depiction <http://www.ex.org/img.png> .

57.
(57)
<span about="http://www.ivan-herman.net/foaf#me">
<span rel="rdfs:seeAlso"
resource="http://www.ivan-herman.net/foaf">
Activity Lead</span>
</span>
 The RDFa @resource attribute is equivalent to
@href
◦ it sets the object, just like @href
◦ but it is ignored by a browser, i.e., not a link!
◦ e.g.,:

58.
(58)
<http://www.w3.org/ns/entailment/RDFS>
dc:creator <http://www.ivan-herman.net/foaf#me> .
<http://www.ivan-herman.net/foaf#me>
foaf:mailbox <mailto:ivan@w3.org> ;
foaf:workplaceHomepage <http://www.w3.org> .
 Here is what we would like to have in RDFa

59.
(59)
<body about="http://www.w3.org/ns/entailment/RDFS">
…
<address>
<span rel="dc:creator"
resource="http://www.ivan-herman.net/foaf#me"/>
<span about="http://www.ivan-herman.net/foaf#me">
<a rel="foaf:mailbox"
href="mailto:ivan@w3.org">ivan@w3.org</a>,
<a rel="foaf:workplaceHomepage"
href="http://www.w3.org">W3C</a>
</span>
</address>
 A straightforward way:

60.
(60)
<body about="http://www.w3.org/ns/entailment/RDFS">
…
<address>
<span rel="dc:creator"
resource="http://www.ivan-herman.net/foaf#me"/>
<span about="http://www.ivan-herman.net/foaf#me">
<a rel="foaf:mailbox"
href="mailto:ivan@w3.org">ivan@w3.org</a>,
<a rel="foaf:workplaceHomepage"
href="http://www.w3.org">W3C</a>
</span>
</address>
 A straightforward way:

61.
(61)
<body about="http://www.w3.org/ns/entailment/RDFS">
…
<address>
<span rel="dc:creator"
resource="http://www.ivan-herman.net/foaf#me">
<a rel="foaf:mailbox"
href="mailto:ivan@w3.org">ivan@w3.org</a>,
<a rel="foaf:workplaceHomepage"
href="http://www.w3.org">W3C</a>
</span>
</address>
 An alternative:

62.
(62)
 @resource (or @href) becomes a subject for the
sub-tree
 This feature is a bit like in RDF/XML

63.
(63)

64.
(64)
 Blank nodes can be created using “_:XX”
 Shorthand for rdf types
 An API has been defined for Web Applications

65.
(65)
 Prefix declarations can be collected in a
separate file and referred to via a @profile
attribute
◦ the “profile file”
◦ RDFa1.0 warning: this is an RDFa1.1 feature!

66.
(66)
 A profile file can also define a term:
◦ assign a URI to a simple string
 The term can be used directly by authors,
without prefixes

67.
(67)
 Say, file “http://ex.org/prof” defines a
mapping:
◦ "desc" → "http://purl.org/dc/terms/description"

68.
(68)
<html prefix="dc: http://purl.org/dc/terms/
rdfs: http://www.w3.org/2000/01/rdf-schema#">
…
<body profile="http://ex.org/prof.html" about="…">
…
<p property="desc">
Unique identifier for <em>RDFS Entailment</em>.</p>
 yields
@prefix dc: <http://purl.org/dc/terms/> .
<…> <http://purl.org/dc/terms/description>
"Unique identifier for RDFS Entailment."

69.
(69)
 Usage of CURIEs and URIs is intuitive for RDF
people…
 It is not for average HTML authors!
 Profile files can be published by major
publishers:
◦ Dublin Core, FOAF, …
◦ FaceBook, Google, …
 … and users can simply refer to the profiles

70.
(70)
 Typing can of course be done using
@rel="rdf:type"
 But that is a widely used combination, so there
is a separate @typeof attribute for that

71.
(71)
<span about="http://www.ivan-herman.net/foaf#me"
typeof="foaf:Person">
<span property="foaf:name">Ivan Herman</span>
</span>,
 yields
<http://www.ivan-herman.net/foaf#me> a foaf:Person ;
foaf:name "Ivan Herman" .

72.
(72)
 RDFa gives an easy way of publishing RDF data
on the Web
 Often, the same RDF data is available in
different formats, including RDFa
◦ it is up to the client to choose which one to use
◦ Web Applications would rely on RDFa, though…

73.
(73)

74.
(74)

75.
(75)
 Various search engines begin to consume RDFa
◦ Google, Yahoo, …
 they may specify which vocabularies they “understand”
 this is still an evolving area
 Facebook’s “social graph” is based on RDFa

76.
(76)
 Embedded metadata (microformat or RDFa) is
used to improve search result page
◦ at the moment only a few vocabularies are recognized,
but that will evolve over the years

77.
(77)
 A number of popular sites publish RDFa as part
of their normal pages:
◦ Tesco, BestBuy, Slideshare, The London Gazette,
Newsweek, MSNBC, O’Reilly Catalog, the White
House…
◦ Creative Commons snippets are in RDFa

78.
(78)Courtesy of Jay Myers, BestBuy, SemTech2010 Presentation

79.
(79)Courtesy of Jay Myers, BestBuy, SemTech2010 Presentation

80.
(80)
 Reported in a BestBuy blog:
◦ GoodRelations+RDFa improved Google rank
tremendously
◦ 30% increase in traffic on BestBuy store pages
◦ Yahoo observers a 15% increase in click-through rate
 Not bad…

81.
(81)

82.
(82)

83.
(83)

84.
(84)

85.
(85)
 Publishing RDFa is an important step in
combining the Semantic Web and the
“traditional” Web
 But publishing is not always straightforward for
a lambda Web designer and user
 This is where the role of Drupal 7 is huge!
◦ make the publication of data in RDFa part of the
normal CMS operation

86.
(86)

87.
(87)
Thank you for your attention!
These slides are also available on the Web:
http://www.w3.org/2010/Talks/RDFa-Drupal-
Tutorial/RDFa.pdf